Hostile Neural Networks guide for All the Mods 10 0.46 to 4.5
Hostile Neural Networks guide for All the Mods 10
Last verified for All the Mods 10 v4.5 on Oct 4, 2026 · text last changed Oct 4, 2026
A newer release (8.3) is out Guide for the current version
Also applies to:All the Mods 10: To the Sky 1.0 to 2.0.2
- Sections
- 8
- Recipes shown
- 15
- Items named
- 19
- Quests covered
- 19
7 sections are written for this pack; 1 is shared from another pack's guide.
Every recipe and number here was checked against All the Mods 10 4.5's own files.
What is Hostile Neural Networks?
2 parts
Hostile Neural Networks lets you simulate mobs and get very real drops out of it. You capture a mob as a data model, train that model, run simulations with it, and then turn the results into mob loot. All the Mods 10 4.5 includes version 6.3.0, and the mod has been in the pack since version 2.25 or earlier.[1][2]
How it works
- Craft a Model Framework and use it on the mob you want. This builds a Data Model for that mob.[1]
- Train the model. Every model starts out Faulty and has to reach Basic in a Deep Learner. The Deep Learner holds a few models at once, and the kills you get train them.[1]
- Once a model is Basic or higher, you can run simulations with it in a Simulation Chamber. Simulations need energy, based on the model type, plus a catalyst, which is usually a Prediction Matrix. Every simulation gives you a generalized prediction. Depending on the model's accuracy, it may also give you a mob-specific prediction. Higher tiers are more accurate, up to 99.5% at Self Aware.[1]
- Feed mob predictions into a Loot Fabricator, which uses power to turn them into drops. You pick a target drop for each model type from a pre-defined list. The fabricator remembers each choice, so one machine can handle several kinds of predictions.[1]
Where it fits in this pack
The pack gives the mod its own quest chapter, and it follows the same flow. It starts with the Deep Learner in a quest in "?" and the Model Framework in a quest in "?". Next comes the Simulation Chamber with a Prediction Matrix in a quest in "?". That quest unlocks the quests for the Generalized Overworld Prediction, Generalized Nether Prediction, Generalized Ender Prediction and Generalized Twilight Prediction, plus the Loot Fabricator in a quest in "?". The Loot Fabricator quest then opens up a long list of quests for individual mob predictions.[3]
Some of these items also show up in another quest chapter, chapter_2_the_star. There, a quest in "?" asks for a Model Framework, and a quest in "?" follows up by asking for a Data Model.[4]
Unverified The sources don't confirm whether Hostile Neural Networks is required to craft the ATM Star, or how important those quests are to that path.
How do you get started with Hostile Neural Networks?
3 parts
Hostile Neural Networks lets you train data models on mobs, run simulations of them, and turn the results into real mob drops. Everything revolves around the Data Model, an item that holds information about one kind of entity. Your first goal is to get one of those and start training it.[1]
Step 1: Model Framework and Deep Learner
The first two items in the pack's Hostile Neural Networks quest chapter are the Model Framework in quest a quest in "?" and the Deep Learner in quest a quest in "?". Both quests sit at the top of the chapter, so craft them first.[3]
Crafting Shaped
| any redstone dust | any redstone dust | |
| any gold ingot |
Recipe id: hostilenetworks:framework
Model Framework
Crafting Shaped
| any obsidian | any obsidian | |
| any glass panes | ||
| any obsidian | any redstone dust | any obsidian |
Recipe id: hostilenetworks:deep_learner
Deep Learner
The Model Framework is how you get a Data Model: use it on the mob you want to target and it builds a model for that entity. Doing this completes quest a quest in "?". The Model Framework and Data Model also appear as quests in the pack's chapter_2_the_star chapter, in quests a quest in "?" and a quest in "?".[1][3][4]
Every new model starts in the Faulty state and has to be trained up to Basic in the Deep Learner. The Deep Learner holds a few data models, and the kills you get train the models inside it. So pop your fresh model in and go fight that mob for a while.[1]
Step 2: Simulation Chamber and Prediction Matrix
Once a model reaches Basic or higher, you can run it in a Simulation Chamber. Simulations need energy, with the amount depending on the model type. They also need a catalyst item, which is usually a Prediction Matrix. The quest a quest in "?" asks for both items and unlocks after the Deep Learner and Model Framework quests.[1][3]
Crafting Shaped
| Empty slot | any glass panes | Empty slot |
| any obsidian | ||
| any lapi gems | any lapi gems |
Recipe id: hostilenetworks:sim_chamber
Simulation Chamber
Crafting Shaped
| any iron ingot | any glass panes | Empty slot |
| any glass panes | any glass panes | |
| Empty slot | any glass panes | any gold ingot |
Recipe id: hostilenetworks:prediction_matrix
Prediction Matrix, the catalyst each simulation uses
Every simulation produces a generalized prediction. It may also produce a Mob Prediction, depending on how accurate your model is. Higher model tiers are more accurate, up to 99.5% at Self Aware, so keep training your models as you go.[1]
The generalized predictions each have their own quest in the chapter, all unlocked by the Simulation Chamber quest. The pack includes recipes that turn a Generalized Overworld Prediction into a Generalized Nether Prediction, and a Nether prediction into a Generalized Ender Prediction.[3][5]
Crafting Shaped
| Empty slot | Empty slot | |
| Empty slot | Empty slot |
Recipe id: hostilenetworks:living_matter/overworld_to_nether
Crafting Shaped
| Empty slot | Empty slot | |
| Empty slot | Empty slot |
Recipe id: hostilenetworks:living_matter/nether_to_ender
Step 3: Loot Fabricator
Mob Predictions become actual drops in the Loot Fabricator, quest a quest in "?". It uses power to convert predictions into mob drops. You pick a target drop from a predefined list for each type of model, and the fabricator remembers your choice per type. That means one fabricator can process several kinds of predictions. Completing this quest opens up the long list of per-mob prediction quests in the chapter.[1][3]
Crafting Shaped
| Empty slot | any netherite ingot | Empty slot |
| any diamond gem | any obsidian | any diamond gem |
| any gold ingot | any gold ingot |
Recipe id: hostilenetworks:loot_fabricator
Loot Fabricator
Once you have more than one fabricator, the Fabrication Directive saves you time. It stores selections from a Loot Fabricator: right-click to copy them, and shift-right-click to apply them to another fabricator.[6]
Crafting Shaped
| Empty slot | any generalized prediction | Empty slot |
| Empty slot | Empty slot |
Recipe id: hostilenetworks:fab_directive
Fabrication Directive
Unverified Exact energy costs for simulations and the Loot Fabricator, and how many kills each model tier needs, could not be confirmed for this pack version.
How does Hostile Neural Networks work?
6 parts
Hostile Neural Networks lets you farm mob drops without farming mobs. You train data models, run simulations with them to produce mob predictions, and then turn those predictions into real loot. Every step of that chain has its own item or machine, so let's walk through it in order.[1]
Data Models: the core of the mod
Everything revolves around the Data Model, an item that holds information about one entity and lets you generate that entity's drops. To get one, you first craft a Model Framework and then use it on the mob you want to build a model of.[1]
Crafting Shaped
| any redstone dust | any redstone dust | |
| any gold ingot |
Recipe id: hostilenetworks:framework
The mod's tooltip text includes a line telling you to click on an entity to attune a model, plus a message saying that model attunement is disabled, so watch the tooltip when you try this.[6]
Which mobs have models is data-driven. Custom Data Models can be added through a datapack, with each model defined as a single file in a data_models subfolder.[1]
Training with the Deep Learner
A fresh model starts in the Faulty state and can't be simulated yet. To level it up, put it in a Deep Learner. The Deep Learner holds a few data models at once, and the kills you get start training them. You need to get each model up to Basic this way before the next machine will accept it.[1]
Crafting Shaped
| any obsidian | any obsidian | |
| any glass panes | ||
| any obsidian | any redstone dust | any obsidian |
Recipe id: hostilenetworks:deep_learner
Running simulations in the Simulation Chamber
Once a model is Basic or higher, you can load it into a Simulation Chamber to run simulations. The Simulation Chamber is the second way to train a model, so your models keep improving while they work.[1]
Crafting Shaped
| Empty slot | any glass panes | Empty slot |
| any obsidian | ||
| any lapi gems | any lapi gems |
Recipe id: hostilenetworks:sim_chamber
Each simulation needs energy and a catalyst item. The energy cost depends on the model type. The catalyst is usually a Prediction Matrix, so you'll want a steady supply of those.[1]
Crafting Shaped
| any iron ingot | any glass panes | Empty slot |
| any glass panes | any glass panes | |
| Empty slot | any glass panes | any gold ingot |
Recipe id: hostilenetworks:prediction_matrix
Every simulation produces a generalized prediction. It may also produce a Mob Prediction, depending on the model's accuracy. Higher model tiers mean higher accuracy, up to 99.5% at the Self Aware tier. Leveling your models is what makes a setup truly efficient.[1]
Generalized predictions
The generalized predictions come in three kinds: Generalized Overworld Prediction, Generalized Nether Prediction and Generalized Ender Prediction. In this pack you can craft them up the chain, turning Overworld into Nether and Nether into Ender.[5]
Crafting Shaped
| Empty slot | Empty slot | |
| Empty slot | Empty slot |
Recipe id: hostilenetworks:living_matter/overworld_to_nether
Crafting Shaped
| Empty slot | Empty slot | |
| Empty slot | Empty slot |
Recipe id: hostilenetworks:living_matter/nether_to_ender
Turning predictions into loot
Mob Predictions become real drops in the Loot Fabricator, a powered device. The possible outputs come from a pre-defined list of items for each model type, and you pick a target for each type. The fabricator remembers your choice per type, so you can feed several kinds of predictions into a single fabricator.[1]
Crafting Shaped
| Empty slot | any netherite ingot | Empty slot |
| any diamond gem | any obsidian | any diamond gem |
| any gold ingot | any gold ingot |
Recipe id: hostilenetworks:loot_fabricator
To reuse your picks, grab a Fabrication Directive. It stores selections from a Loot Fabricator: right-click to copy, shift-right-click to apply.[6]
Crafting Shaped
| Empty slot | any generalized prediction | Empty slot |
| Empty slot | Empty slot |
Recipe id: hostilenetworks:fab_directive
What the tooltips tell you
The mod has several info lines that are useful for planning:[6]
- A prompt to hold a key for more info, so check the extended tooltip.[6]
- Generates Experience, followed by a value.[6]
- Loot Table Copies, followed by a value.[6]
- No Base Drop Produced.[6]
Unverified The sources don't give exact energy costs, how much power the Loot Fabricator uses, the kill or simulation counts needed for each tier, the names of the tiers between Basic and Self Aware, or which sides of the machines accept items and energy for automation. Check these in-game before you design an automated setup.
How do you progress through Hostile Neural Networks?
4 parts
The Hostile Neural Networks questline in this pack is a simple branching tree. You start with two quests side by side, combine them into your core machines, then branch into generalized predictions and the Loot Fabricator. Here's the path, step by step.[3]
Step 1: Deep Learner and Model Framework
The chapter opens with two parallel quests. a quest in "?" asks for the Deep Learner, and a quest in "?" asks for the Model Framework. You need both before the rest of the chapter opens up, so craft them together.[3]
Crafting Shaped
| any obsidian | any obsidian | |
| any glass panes | ||
| any obsidian | any redstone dust | any obsidian |
Recipe id: hostilenetworks:deep_learner
Crafting Shaped
| any redstone dust | any redstone dust | |
| any gold ingot |
Recipe id: hostilenetworks:framework
Both of these use the mod's own recipes in this pack, not pack-modified versions. The same goes for every other Hostile Neural Networks recipe on this page.[5]
Step 2: Your first data model and the core machines
Finishing both starter quests unlocks two more. a quest in "?" asks you to get a data model. a quest in "?" asks for both the Simulation Chamber and the Prediction Matrix in a single quest.[3]
Crafting Shaped
| Empty slot | any glass panes | Empty slot |
| any obsidian | ||
| any lapi gems | any lapi gems |
Recipe id: hostilenetworks:sim_chamber
Crafting Shaped
| any iron ingot | any glass panes | Empty slot |
| any glass panes | any glass panes | |
| Empty slot | any glass panes | any gold ingot |
Recipe id: hostilenetworks:prediction_matrix
The Simulation Chamber and Prediction Matrix quest is the hub of the whole chapter. Every later quest hangs off it, so it's worth prioritising.[3]
Step 3: Generalized predictions
Once the hub quest is done, four quests open, one for each generalized prediction.[3]
- a quest in "?" asks for the Generalized Overworld Prediction.[3]
- a quest in "?" asks for the Generalized Nether Prediction.[3]
- a quest in "?" asks for the Generalized Ender Prediction.[3]
- a quest in "?" asks for the Generalized Twilight Prediction.[3]
The pack has two crafting recipes that step generalized predictions up a dimension. One turns Overworld predictions into the Generalized Nether Prediction. The other turns Nether predictions into the Generalized Ender Prediction. If you have plenty of one kind, you can climb the chain instead of farming each one separately.[5]
Crafting Shaped
| Empty slot | Empty slot | |
| Empty slot | Empty slot |
Recipe id: hostilenetworks:living_matter/overworld_to_nether
Overworld to Nether
Crafting Shaped
| Empty slot | Empty slot | |
| Empty slot | Empty slot |
Recipe id: hostilenetworks:living_matter/nether_to_ender
Nether to Ender
Unverified The pack's recipe list has no crafting recipe that produces the Generalized Twilight Prediction or the Generalized Overworld Prediction. They presumably come from running data models, but the sources don't confirm this.
Step 4: Loot Fabricator and mob predictions
The last branch off the hub is a quest in "?", which asks for the Loot Fabricator. This is the biggest unlock in the chapter: completing it opens a long list of quests, each asking for a prediction item, such as a quest in "?", a quest in "?" and a quest in "?".[3]
Crafting Shaped
| Empty slot | any netherite ingot | Empty slot |
| any diamond gem | any obsidian | any diamond gem |
| any gold ingot | any gold ingot |
Recipe id: hostilenetworks:loot_fabricator
You can also craft the Fabrication Directive. It isn't a quest task in this chapter, but its recipe uses a Prediction Matrix and any generalized prediction, so you'll be able to make it once you reach this stage.[3][5]
Crafting Shaped
| Empty slot | any generalized prediction | Empty slot |
| Empty slot | Empty slot |
Recipe id: hostilenetworks:fab_directive
Unverified The data model tier names, how much data each tier needs, and what the Simulation Chamber and Loot Fabricator need to run aren't covered by the sources, so they're left out here.
How does Hostile Neural Networks work with other mods?
2 parts
The biggest cross-mod link for Hostile Neural Networks in this pack is with Twilight Forest. The pack includes shapeless crafting recipes that turn a Generalized Twilight Prediction into Twilight Forest materials, so your simulations can supply Twilight resources without farming them by hand.[5]
Twilight Forest materials from predictions
Each recipe gives you a different Twilight Forest item. Here's what you can make, with the exact in-pack recipe shown for each one.[5]
Crafting Shapeless
Generalized Twilight Prediction
Stick
Recipe id: hostilenetworks:living_matter/twilight/liveroot
Crafting Shapeless
Recipe id: hostilenetworks:living_matter/twilight/raven_feather
Crafting Shapeless
- any leave
Recipe id: hostilenetworks:living_matter/twilight/steeleaf
Crafting Shapeless
Recipe id: hostilenetworks:living_matter/twilight/torchberries
Generalized predictions come out of the Simulation Chamber: every simulation you run always produces a generalized prediction, alongside a chance at a mob-specific one.[1]
The mod's own questline has a quest for the Generalized Twilight Prediction, a quest in "?". It unlocks after the quest for the Simulation Chamber and Prediction Matrix, a quest in "?".[3]
Unverified Which data models produce the Generalized Twilight Prediction in this pack is not confirmed by the available sources.
Hostile Neural Networks in the main progression
Hostile Neural Networks items also show up in quests outside the mod's own chapter. The pack's chapter_2_the_star chapter asks you for a Model Framework in a quest in "?", then for a Data Model in a quest in "?", which comes after it. So building your first data model is part of the pack's wider progression, not just an optional side project.[4]
Which All the Mods 10 quests use Hostile Neural Networks?
4 parts
Hostile Neural Networks has its own quest chapter in this pack, hostile_neural_networks, and it also shows up briefly in a second chapter, chapter_2_the_star. The main chapter follows a clear order: first the basic tools, then the machines, then a big batch of Prediction collection quests.[3][4]
Main chapter: getting started
The chapter opens with two side-by-side quests that both unlock after an earlier quest in the chapter. a quest in "?" asks for a Deep Learner, and a quest in "?" asks for a Model Framework. You can do them in either order.[3]
Once you've finished both, two more quests open up. a quest in "?" asks for a Simulation Chamber and a Prediction Matrix. a quest in "?" asks you to get a Data Model.[3]
Branching out from the Simulation Chamber
The Simulation Chamber quest is the big hub of the chapter. Completing it unlocks five quests at once, and you can tackle them in any order:[3]
- a quest in "?": get a Generalized Overworld Prediction.[3]
- a quest in "?": get a Generalized Nether Prediction.[3]
- a quest in "?": get a Generalized Ender Prediction.[3]
- a quest in "?": get a Generalized Twilight Prediction.[3]
- a quest in "?": build a Loot Fabricator.[3]
The Prediction collection
The Loot Fabricator quest unlocks the largest part of the chapter: a long set of quests that each ask for a Prediction. They all have the Loot Fabricator quest as their only prerequisite, so they open together and you can finish them in whatever order your data models allow. Examples include a quest in "?", a quest in "?", a quest in "?", a quest in "?", a quest in "?", a quest in "?", a quest in "?" and a quest in "?", and there are many more in the same group.[3]
Unverified The quest data only shows the generic Prediction item for each of these quests, so it's not confirmed which mob's Prediction each one asks for. Check each quest in-game to see what it needs.
Appearance in another chapter
The chapter_2_the_star chapter also has two quests for this mod. a quest in "?" asks for a Model Framework, and after that a quest in "?" asks for a Data Model. These are the same items the main chapter asks for early on, so you'll probably make progress on both chapters at once.[3][4]
Which Hostile Neural Networks config options matter?
Same as in the All the Mods 10 8.1 guide
The mod's own project page describes a few numbers and costs you'll deal with most, but it doesn't list any config options. It says running a simulation in the Simulation Chamber costs energy, and the amount depends on the model type. Each simulation also uses up a catalyst item, usually a Prediction Matrix. It also says accuracy goes up as your Data Model reaches higher tiers, topping out at 99.5% at Self Aware.[1]
Keep in mind that this is how the mod describes itself in general, not a confirmed All the Mods 10 8.1 setting. The pack could use different values.[1]
Unverified We couldn't check this pack's Hostile Neural Networks config values, so they aren't listed here. That means we can't tell you whether the pack changed any mod defaults, such as simulation energy costs, accuracy per model tier, or Deep Learner and Simulation Chamber settings. We also couldn't confirm which of these are configurable in the first place. To see the exact values, open the mod's config file in your own All the Mods 10 8.1 instance.
What changed in Hostile Neural Networks between versions?
The sources I have for Hostile Neural Networks only cover version bumps in the pack, so this is a short history. There are no stored changelogs for the shipped jars, which means I can't tell you what actually changed inside each mod update.[2][7][8]
- All the Mods 10 2.2 or earlier: Hostile Neural Networks was already part of the pack by this point.[2]
- All the Mods 10 2.47: the mod was updated from 6.1.3 to 6.2.0.[8]
- All the Mods 10 4.3: the mod was updated from 6.2.0 to 6.3.0.[7]
- All the Mods 10 4.5: the pack ships version 6.3.0 (the 1.21.1 build), so there has been no change to the mod since 4.3.[2][7]
Unverified What changed in the mod between 6.1.3, 6.2.0 and 6.3.0 (new features, balance changes, fixes) is not confirmed by the available sources, and neither is whether the pack altered any of its recipes or config values in those releases.
Sources
- Hostile Neural Networks project page (modrinth) · official mod data
- All the Mods 10 4.5 mod list · pack file
- All the Mods 10 4.5 quests: "hostile_neural_networks" · pack file
- All the Mods 10 4.5 quests: "chapter_2_the_star" · pack file
- All the Mods 10 4.5 recipe dump: recipes making Hostile Neural Networks items · pack recipe dump
- hostilenetworks tooltips and descriptions (English lang file) · official mod data
- All the Mods 10 changelog (All the Mods 10 4.3) · changelog
- All the Mods 10 changelog (All the Mods 10 2.47) · changelog
Where this guide comes from
Written from what All the Mods 10 itself ships: its files and recipes, the in-game guidebooks and the mods' own pages. Every statement links to the source it came from, and the numbers are checked against the pack before publishing.
Guide last verified Oct 4, 2026.
Guides for other versions: 4.6 to 8.3 · 0.45